Wheel Expressions Getting The Object Space .tx

[george-stankevich]

Hello everybody:

I am having a problem with an automatic wheel rotation script for a car. The expression works fine at first, but when the car rotates (changes direction) the expression breaks. I wrote the expression to use the TranslateX to drive the Y rotation of the wheels. It seems that when the car turns 90 degrees the Z axis is now the front axis and the wheels no longer rotate.

Is there any way to get an object space rotation from the channel box.

I suppose I could also use xform but when I use Mel syntax the expression it will not evaluate unless I set keyframes. Worse it does not evaluate at render time.

George

[joojaa]

QUOTE(xenondrink @ 07/16/08, 01:57 PM) [snapback]288408[/snapback]
I am having a problem with an automatic wheel rotation script for a car. The expression works fine at first, but when the car rotates (changes direction) the expression breaks. I wrote the expression to use the TranslateX to drive the Y rotation of the wheels. It seems that when the car turns 90 degrees the Z axis is now the front axis and the wheels no longer rotate.

Well yes for your questions but that alone does not help.

A location is a vector. Rotation of a wheel is a function of the line integral against the coefficent of the direction of the wheel. Thats a nifty way of saying its the distance of the path traveled so far against how the wheel happened to be aligned at every possible moment to that path.

See the problem is that a wheel can reach any point of space with a unlimitted amount of solutions. So at any given point in space the wheel can have any rotation. Now the rpblem is that all this has NO perfect solutions, just sveral problematic ones.

So no theres no value you can ask fr free movement. You could compute it tough no problem. Easiest schenario is to just do a per frame calculation of path length. This offcourse DOES NOT work out in realtime it works out in playback when every frame is played. It still has a bit of problem, deciding wether or not the wheel rotates forward or backward is a bit tricky (ditances are allways positive), also knowing when the wheel follows a hill upwards and when its flying and does not need to rotate right si a bit of a problem, like is any slipage etc. (this is by far the most used one tough i dont really care so much for this solution anymore)

If however yu use patth anim you can just ask the value travelled, tough for a 4 wheeled wehicle its not perfectly accurate as the wheel on the inner of a curve would need t rotate less.

Get some of the premade car rigs.

So not as easy as you think Not terribly hard to guestimate but then again it depends on your level of profficency. I am a mechanical engineer by education so stuff like this was part of what i was taught out. Thing si that the full working of a car wheel is REALLY complex, took about 4 years to teach out all thats needed for that after all basic education including college. And i still cant quite figure out all the nuances.

[george-stankevich]

Thank You

I am thinking mabey I could impliment your idea by using a motion path and attaching the wheel's rotation to the U value of the path. This would not take care of the flying bit, unless I subtracted the ammount of path traveled while the car is not touching the path from the uvalue to create a new uvalue - flying distance * rotation multiplier = wheel rotation type equation.

I had not even thought about the rotation of the wheel going around a corner. I suppose the best thing might be to add an manual rotation offset that allow the user to slow or speed up the automatic rotation by adding in their own rotation offset.

This would give the animator a basic rotation and flexibility, it would also eliminate the need the troublesome flying distance offset mentioned earlier.

At this point I have a solution that uses "xform -q -t -os carName" to allow the wheels to rotate along the local axis even when the car is turned. The standard "wheel.rx = car.tx * $multValue" breaks when the car is turned. The problem with this new setup is that the wheels go crazy when the car is being rotated, after the rotation it is fine but durring the rotation the wheel jumps around unperdictably.

Thanks again,

xenondrink

[calexbee]

Hey I just had that similar problem and I found an awesome tutorial that fixes your wheel rotation and direction issue. Here's the link.

http://www.lostpencil.com/wordpress/?p=231

[joojaa]

Thet would be the fisrt method yes, with flip control still does not account all scenarios one would typically need. If i would do motion path anim i wouldn't use this its a bit fo waste of tie. Ok if you have less than 50 wheels sure.

BUt aslo it does not work all that well with inclination. And that sucks. Because half of the time youd use 3 dims anyway. Making teh 3rd dimesnion isnt that hard tough, yoou get slippage/turning for free too if you wanted.

Not really in deep maya knowledge here! Id say this is normal rigging skill level.